Light-activation of DNA-methyltransferases

Abstract

The most prominent DNA modification in mammalian cells is the 5-methylation of cytosine (5mC) in a CpG context and several severe diseases like cancer are connected to aberrant cytosine methylation. 5mC is a dynamic key epigenetic modification regulating chromatin states and therefore gene transcription, and it is highly important to get new insights into the regulation and effects of such epigenetic mark. It is necessary to be able to write 5mC at user-defined genomic loci at specific time points to analyze locus specific and time-resolved downstream effects. 5mC is written by DNA methyltransferases (DNMTs) and several approaches were made to control 5mC levels either on a global or targeted gene level. However, these approaches had low spatial and/or temporal control and resulted in off-target effects. For now, it was not possible to control the catalytic activity of DNMTs itself. This work reports the direct light-control of in vivo DNMT activity to overcome drawbacks of previous approaches. A photocaged cysteine is genetically encoded on DNMT3a to replace the Cys710 which is essential for the catalytic activity. This leads to the expression of DNMT3a proteins in an inactive state and simultaneous activation of them is achieved by light irradiation. This results in the cleavage of the caging group and the remaining of an unmodified and active DNMT3a. Since upstream processes like transfection and protein expression are uncoupled from the actual catalytic activity after light-activation, the kinetics of 5mC writing alone can be monitored. This tool is used to study the effects of several DNMT3a mutations connected to acute myeloid leukemia on the catalytic activity, providing new in vivo observations to dissect the role of these mutations. In addition, it is a great advantage to be able to write 5mC at user-defined genomic loci to specifically alter the chromatin state or to monitor protein interactions associated with 5mC at such loci. Here, locus-specificity is reached by fusing a recombinant DNMT protein with a programmable transcription activator like effector protein. Thereby, inactive DNMTs are recruited to the target locus and activated at a given time point which prevents off-target methylation. It is noteworthy that the final 5mC level is tunable by adjusting the light irradiation time. Also, time-resolved effects of DNA methylation on the transcriptome are reported, providing data on the fast consequence of DNA methylation on gene expression. In conclusion, the ability to control the activity of DNMTs with least structural change by just incorporating a single non-canonical amino acid and the rapid activation of such by light gives previous unreachable spatio-temporal control over DNA methylation in living cells. With this tool, the kinetics of downstream effects of 5mC can be monitored and new insights into the 5mC-related epigenetic network can be obtained